SOCRATES’S AVOWAL OF KNOWLEDGE REVISITED

The sense in which Socrates avows or disavows knowledge has beenmuch debated in treatments of Plato’s epistemology over the last coupleof decades. This paper aims to outline a possible interpretation thatscholars of Socrates have yet to elaborate fully; it seeks to shed lightespecially on the Socratic method of enquiry and on the nature ofSocratic knowledge at 29 b in Plato’s Apology, i. e. its shareability andthe content- and context-dependent nature of his claim to it.

Introduction

Socrates’s avowals have attracted a range of interpretations, includ-ing the theses that [1] Socrates’s disavowal of knowledge is best under-stood as an expedient inviting his interlocutors to search with himjointly for the truth; [2] Socrates’s repeated disavowal of knowledgeshould be taken seriously, and his avowals of knowledge regarded asexceptional; somewhere in between, [3] in avowing and disavowingknowledge, Socrates uses two senses of the word ‘know’; [4] Socratessometimes avows knowledge in Plato’s earlier dialogues, but hardly atall in the middle or later dialogues, suggesting that we can read Plato’sthought as developmentalist or revisionist in epistemic terms; and [5]there may be some reason that Socrates is inconsistent in his avowalsand disavowals, e. g. he has some pedagogical purposes in mind. All ofthese interpretations have their own weaknesses – referring to the num-bering above, I introduce my grounds for reservations at selected pointsin the body of my discussion. While engaging with these central read-ings, this essay primarily develops an independent line of argument con-cerned with the grounds for considering Socrates’s avowals qualified.

1. The disavowal and avowal of knowledge

Plato’s Apology, in common with other early Platonic dialogues,impressively describes what may be termed the negative side of Socrates’sepistemic state, namely his disavowal of knowledge.1 Socrates calls the

1 It may be worth noting, with M. C. Stokes (ed.), Plato, Apology (Warminster1997) 99, 195, that the Apology begins and ends with a statement of the negativity

035_056_Ohkusa.pmd 11.05.09, 13:3335

Hyperboreus Vol.14 (2008) Fasc.2

36 Terumasa Ohkusa

pronouncement of the Delphic oracle that none is wiser than he anenigma, since he does not acknowledge himself to be wise in any mat-ter, great or small (21 b). But Socrates does admit some light and shadein the question of his wisdom or ignorance; he is wiser at least than apolitician who had professed knowledge, but who on examinationturned out to know nothing fine and good (21 d 3–8).2

Some understand Socrates’s disclaimer of knowledge as a mere tac-tic, encouraging his interlocutor to seek out the truth (interpretation[1]).3 Socrates, however, appears sincere in insisting upon his lack ofknowledge at 23 a 3–7. In gesturing towards divine wisdom here,Socrates defers, in whatever human wisdom he may be supposed tohave, to a greater authority.

But what if we still refuse to take this Socratic formulation at facevalue? The passage at 23 a cannot provide in its own right a full guaran-tee that Socrates is genuine in positing his own ignorance, as he doesover the course of the Apology. That is, his explanation might again betaken to be question-begging in that it could be construed as merely afurther expedient to secure some objective. We could doubt Socrates’sgenuineness in disavowing knowledge indefinitely in this way – alwayssupposing that he has some reason to conceal a positive epistemic state.

My claim is that, in assessing the sincerity attaching to Socrates’sstatements, we should consider the immediate discursive context of theApology, where Socrates is standing trial. This means giving due weightto a series of statements he makes in court. The courtroom context is notone where Socrates can say things lightly, both because of the gravity ofthe charges against him, and because he expects that the jury’s verdictwill conform to general standards of truth and justice (18 a 3–6). Impor-tantly, Socrates claims at the beginning of the Apology that the casepresented by his accusers is quite devoid of truth, whereas his will de-

of Socratic cognition (... oÙk o�da, 17 a 1�2; ¥dhlon pant� pl¾n À tù qeù, 42 a3–5). This is especially striking when one considers the natural tendency of trialsto urge their defendants to make positive statements. On the Apology as a historicaldocument, see Guthrie’s careful treatment (W. K. C. Guthrie, A History of GreekPhilosophy IV [Cambridge 1975] 72–80). Guthrie attempts to reconstruct thehistorical Socrates, whose professions of ignorance he does not judge to be in anyserious sense deceitful or “insincere” (ibid., III [Cambridge 1969] 447). Myconcern is not with determining the pronouncements of a historical Socrates, butwith considering the totality and the possible consistency of the statementsregarding knowledge expressed by “Socrates” in Plato’s Apology. My reference tothe Apology is from the new Oxford Classical Text (Platonis Opera I [Oxford1995]) edited by E. A. Duke et al.

2 Cf. n. 24.3 N. Gulley, The Philosophy of Socrates (New York 1968) 64–69.

035_056_Ohkusa.pmd 11.05.09, 13:3336

37Socrates’s Avowal of Knowledge Revisited

liver up to the bystanders the truth in its entirety (17 b 7–8).4 Socrates’spositioning of his denials of knowledge within the context of an evi-dently serious appeal to truth gives us good reason for supposing hisdisavowals to be literal.

Moreover, the characteristic procedure of Socratic dialogue itselfprovides a basis for acquitting Socrates of mere rhetoric in his disclaim-ers of knowledge. Let us accept the position suspicious of Socrates thatit is not enough for him to keep pressing the truth of his assertions. Inthat case, the validity of both sides’ arguments will depend on their hav-ing a solid foundation in fact (17 b 2, �rgJ; 32 a 5, �rga etc.). We mustthen examine how Socrates seeks to establish his proofs. He charac-teristically proceeds through refutation of the claims of his accusers,rather than through positive statements in their own right. This proce-dure describes the typical Socratic elenchus, a mode of argumentwhich, I shall suggest, bears an essential relation to Socrates’s episte-mological claims. In court, Socrates tries to refute the problematicpoints of each charge in turn, just as he does in his ordinary discourse.Socrates takes care not to make one-sided statements, but rather, in re-sponse to persistent accusations, demands from his bystanders there andthen a discussion of whether these are based on facts (19 d 1–7).Socrates’s cross-examination of one of his accusers, Meletus, in dealingwith the later allegations, offers an example of his argumentative style(24 c 10 – 28 a 2). Opening his defence in a style of homespun diction,after his accusers had warned the audience of the delusive persuasive-ness of his skilful speech, Socrates bases his whole argument on amethod of elenctic disproof which resorts to counter-evidence at everyopportunity (cf. 19 d 1–2, 20 e 5 – 21 a 9, 24 a 7–8, 31 c 2–3, 32 a 4–5,32 d 1, 32 e 1, 33 d – 34 a, etc.). Insofar as Socrates seems to depend forhis argument on the consent he gains from the audience, we cannot thenregard his repeated use of the word ‘truth’ as groundless.5

4 This strong contrast between Socrates and his accusers in terms of truth-fulness is made in especially concentrated and vigorous fashion in the Proo�mion(17 a 1 – 18 a 6), which ends with the phrases t¢lhqÁ l�gein (cf. J. Burnet [ed.],Plato’s Euthyphro, Apology of Socrates and Crito [Oxford 1924] 153), and fromthen on appears repeatedly in the Apology. On the truthfulness of the case forSocrates, see 17 b 4–5, 8, 18 a 6, 20 d 5–6, 22 b 6, 24 a 5–8, 28 a 7, d 6, 29 a 2, d8 – e 3, 31 c 2, e 2, 32 a 8, 33 c 2, 9, 34 b 6, 39 b 5, 41 c 9 etc. On the falsity of theindictment, see 17 a 3–5, b 7, 18 a 8, b 2, 20 e 3, 26 a 1, 27 e 5, 30 b 6, 33 b 8, 34b 5, etc.

5 Socrates seems to pay special attention to his prosecutors’ reactions over thecourse of his speech (e. g. 27 c 10). In contrast, his prosecutors submit neitherevidence nor witnesses whose testimony may be confirmed by both sides, even

035_056_Ohkusa.pmd 11.05.09, 13:3337

38 Terumasa Ohkusa

We can take his disavowals at face value, however, and be no furtherforward in understanding them. Socrates’s acknowledgment of his lackof knowledge suggests that a certain cognitive state should be attributedto him.6 In practice, however, Socrates expresses a number of positiveviews, particularly with regard to morals.7 The question then becomes:does Socrates, in expressing some positive views, make some profes-sion tantamount to an avowal of knowledge?

As commentary has acknowledged,8 Socrates makes assertions ofknowledge a number of times in the Apology, especially perhaps inspheres other than moral philosophy.9 While not exactly numerous,these assertions are more than enough for us to show that Socratesavows knowledge at certain junctures. With regard to the possibility ofmoral knowledge, on the other hand, Socrates in Plato’s early dialoguesseems to lay down a stringent condition: in order for a person to avowknowledge of something, he needs to answer the Socratic question whatit is.10 When examining his interlocutors, however, Socrates does notneed to be able to answer this definitional question himself in order to

when Socrates encourages them to do so (cf. 31 b 8 – c 1, 34 a 3–7). The accusers’one-sided allegations turn out to lack substantial support and also to be riven withself-contradiction, e. g. Meletus (M�lhtoj) – good and patriotic as he purports tobe – turns out never to have cared (oÙd�n mem�lhken, cf. 24 c 4–9, d 9–10, 25 c 1–4, 26 a 9 – b 2) about the education of the young.

6 On a potential problem in claiming to recognise one’s own ignorance, seen. 28.

7 Socrates claims never to have yielded to anyone in anything contrary tojustice (33 a) and intends never to do so (32 a, cf. 37 a, b). He posits that thegreatest good for a human being consists in holding daily discussions about virtueand other topics; the unexamined life is not worth living (38 a). For Socrates, it isshameful to concern oneself with the acquisition of as much money, repute andhonour as possible, instead of fixing one’s mind on prudence, truth and the bestpossible condition of one’s soul (29 e).

8 J. H. Lesher, “Socrates’ Disavowal of Knowledge”, Journal of the Historyof Philosophy 25 (1987) 280; C. D. C. Reeve, Socrates in the Apology (Indiana-polis 1989) 54–55; H. Benson, Socratic Wisdom (Oxford 2000) 222–226, etc.

9 Cf. 21 b 5, 22 b 8, d 1, 24 a 6, 25 d 9 – e 4 (in the form of a rhetoricalquestion), 37 b 7, d 6. What I am stressing here as “morals” are subjects whichSocrates takes up as central issues in Plato’s early dialogues (justice, temperance,courage, piety and wisdom). Despite the implication of several passages likeEuthphr. 7 b–c or Phdr. 263 a–b, however, it is problematic – or at least difficult –stringently to demarcate the realm of morals and that of other topics (cf. La. 187e – 188 a, Grg. 490 b – 491 a, 497 b–c; note the usage of t¦ smikr£ for both ‘thepetty trifles’ and ‘the lesser mysteries’). At this point in our treatment of the Apo-logy, what should be noted is how often Socrates avows knowledge. For qualificationsof Socrates’s claim to knowledge, see also 18 d 1, 20 e 8, etc.

10 Cf. La. 190 c 6, Euthphr. 5 c 4 – d 1, Hp. Ma. 286 c 8 – d 2.

035_056_Ohkusa.pmd 11.05.09, 13:3338

39Socrates’s Avowal of Knowledge Revisited

expose his partner’s avowal of knowledge as deluded. Thus Socrates’sprocedure entails no affirmation of positive knowledge. Even in thecase of the oracle’s endorsement of his wisdom, the conclusion ad-vanced by Socrates was that the name “Socrates” was “additionally”used by the god to represent anyone, using the person Socrates as anexample, who recognises his own lack of knowledge (23 a–b). SinceSocrates in his interpretation of the oracle estimates human wisdom tobe worth little or nothing, the context in which such wisdom is bestdeployed is one in which the subject of “knowledge” accepts its finitudeor inadequacy.

This representation of Socratic ignorance follows the interpretationthat attributes to Socrates a thoroughly negative epistemic attitude.Irwin, for instance, argues that Socrates disclaims all knowledge, ap-pealing to a clear distinction between knowledge and the true beliefwhich admits of positive conviction.11 However, this line of interpreta-tion appears irreconcilable with Socrates’s definitive statement that heknows that “to do injustice, that is, to disobey the superior, god or man,is bad and shameful” (29 b 6–7).12 The gravity of his insistence isunmissable. The word o�da here falls with especial emphasis at the endof a periodic sentence, insisting upon Socrates’s claim to knowledge.Further, this avowal of knowledge is all the more striking, since it con-trasts with Socrates’s immediately preceding disclaimer of knowledgeof life after death. Here, Socrates says that many people fear deathwrongly, thinking that they know what they do not know; in summarythis attitude is based on a blameworthy ignorance (¢maq�a ... �pone�-distoj, 29 a–b), which has the same structure as that previously diag-nosed in the politician by Socrates.13 While the English ‘ignorance’may not convey the exact nuance of this Socratic use of the Greek¢maq�a, Socrates’s intention is clear: to point up people’s delusion orfalse conceit of knowledge. Having traced the root of the fear of death in

11 T. Irwin, Plato’s Moral Theory (Oxford 1977) 58 suggests that Socrates“exceptionally” claims knowledge at Ap. 29 b 6–7 (interpretation [2]). However,Irwin deals with 29 b a bit differently in Plato’s Ethics (Oxford 1995) 27–29,where he sees o�da in question as Socrates’s conviction, not knowledge. On thischange, cf. G. Fine, “Nozick’s Socrates”, Phronesis 41 (1996) 234–235 n. 6.

12 Cf. E. de Strycker, S. R. Slings, Plato’s Apology of Socrates (Leiden 1994)327 and oÙk o�da in my n. 1.

13 On such delusion, often described as a peculiar type of ignorance by Plato,see Lg. 863 c. This type of cognitive state, distinguished from the empty form ofnot knowing, and often paraphrased as ¢maq�a, recurs throughout Plato’s works,cf. Ap. 21 c–d, 22 b–e, 23 c–d, 29 a–b, e, 41 b, Alc. I. 117 d – 118 b, Chrm. 166 d1–2, 167 a 4–5, Ly. 218 a–b, Men. 84 a–c, Smp. 204 a, Phdr. 275 b, Tht. 210 c, Sph.229 c, Plt. 302 a–b, Phlb. 48 c – 49 a, Lg. 732 a–b, etc.

035_056_Ohkusa.pmd 11.05.09, 13:3339

40 Terumasa Ohkusa

this way, he does not regard the knowledge expressed at 29 b as liable tosuch delusion.14 The knowledge at 29 b, rather, works as a guide for himin decision-making (29 b 8 – c 1).15

2. Knowledge at 29 b

At this point, it is necessary to seek further to characterise the natureof the knowledge that Socrates seems to claim. Vlastos distinguishesbetween two conceptions of knowledge: strong and weak (interpreta-tion [3]).16 For Vlastos, the criterion distinguishing these differentkinds of knowledge concerns the certainty with which they may beheld. Strong knowledge is attributed only to the god and carries with itinfallible certainty. Weak knowledge may, on the other hand, be arrivedat through the Socratic method of elenchus, and remains fallible anduncertain. Agreements which have survived thousands of interrogatoryand testing processes of elenchus may well be overturned at the nextchallenge. Indeed, the introduction of two types of knowledge seems

14 This line of argument concurs with R. Kraut, Socrates and the State (Princeton1984) 275: “in the Apology he [Socrates] confidently and deliberately claims toknow” at 29 b. Benson (n. 8) 238 confusingly describes Socrates’s assertion as a“misstatement made in the heat of the moment”, while also drawing attention to “acertain amount of deliberateness” in Socrates’s statement at 29 b (ibid., 236).

15 Socrates relates knowledge of such a kind at 37 b 7, with an emphatic form:eâ o�da. This is possibly the passage to which Cicero refers by his term ‘ante’ atTusculanae disuputationes I 99, where Cicero highlights a knowledge on Socrates’spart that would allow him to choose between 1) pursuing an examined life even atrisk of death, and 2) abandoning philosophy and begging for his life. Cicero hadpreviously been summarizing 40 c – 42 a of Plato’s Apology. 29 b is anothercandidate for Cicero’s reference.

16 Cf. G. Vlastos, Socratic Studies (Cambridge 1994) 39–66; idem, “Socrates’Disavowal of Knowledge”, Philosophical Quarterly 35 (1985) 1–31. The intro-duction of determinately two different sorts of knowledge has no firm basis inPlato’s texts and further risks too much schematization, in a context where it isotherwise only natural to assume a range of differences in knowledge’s content asI will discuss in the body of this chapter.

I agree with T. C. Brickhouse & N. D. Smith, Plato’s Socrates (Oxford 1994)38–45 that Socrates is interested in knowledge-how (or why) rather than know-ledge-that: there is textual evidence for this distinction in Ópwj �cei at 509 a 5 inthe Gorgias; and indeed other references, such as Chrm. 166 d 6 ÓpV �cei, Euthd.278 b 5 t¦ pr£gmata ... pÍ �cei, Euthphr. 4 a 12, e 5 ÓpV �cei, Men. 84 b 10 ÓpV�cei, Cra. 420 b 5 ÓpV �cei ... t¦ pr£gmata etc., back up a distinction forSocrates in kinds of knowledge. But in my argument, this distinction betweenknowledge-how and knowledge-that is assimilated to the suggestion that thecontent of a knowledge-claim is of greater epistemological moment than themeaning of the word ‘knowledge’ that conveys it.

035_056_Ohkusa.pmd 11.05.09, 13:3440

41Socrates’s Avowal of Knowledge Revisited

reasonable given that Socrates both denies and claims some knowledge.We are now in a position to enquire whether the avowal at 29 b, whichSocrates notably contrasts with people’s delusion or false conceit ofknowledge, ¢maq�a, merely represents an assertion of this fallible, un-certain cognition.

For Socrates, any assertion of knowledge obliges speakers un-complacently to articulate their thinking and to defend themselves inpublic. Knowledge must be capable of being shared and tested in thecontext of a dialogue between two interlocutors.17 To take the case ofthe knowledge asserted at 29 b, it is presumably crucial for Socrates’sstatement – “for one to do injustice, that is, to disobey his superior, godor man, is bad and shameful” – that it is clear and readily shareable withhis accusers. In a pragmatic sense, the statement is shared in that itforms the foundation for the trial in which Socrates is being accused ofimpiety. Socrates appears to treat the content of his statement about in-justice as uncontroversial in the courtroom context of his speech.

If so, is the claim of knowledge at 29 b by Socrates context-depen-dent? We may well say yes, so long as the question turns upon whetherSocrates claims knowledge or not. By and large, the situations in whichwe claim to know something place large restrictions upon the nature ofour claims. Within ordinary language, it is unusual for the content ofour statements to be simply that we know something; this is only usuallythe case in junctures, such as Socrates’s in the Apology, where aperson’s claims to knowledge are doubted. In other words, certain dis-cursive contexts (say, when we are embattled) remove from us the pre-condition that our statements are the object of previous consent. Some-times, when consent to specific assertions turns out to be controversial,the status of knowledge itself needs re-examining.18 This is a fortiorithe case with Socrates, who is notably cautious and uncomplacent inargument. Socrates open-mindedly examines any issue from first prin-ciples when facing a proponent of an opposite value-judgment, whomight contest a basic statement like 29 b.19 It should be noted here,though, that Socrates’s willingness to examine without prejudice the

17 Cf. Chrm. 166 c 7– d 4.18 Cf. Euthphr. 7 b–d. Usually, though, knowledge-claims about assertions,

i. e. that in asserting x, we also assert “I know x”, are not something that needremarking on in Socrates’s dialogues.

19 Socrates professes amazement when Thrasymachus counts injustice as akind of excellence and wisdom and justice as defects (cf. R. I, 348 c – 349 a).Although to some degree, Socrates understands Thrasymachus’s contention thatprofit may flow from wrongdoing, Thrasymachus’s strong position that injustice“is fine (kalÒn) and strong” makes it difficult for Socrates to determine the

035_056_Ohkusa.pmd 11.05.09, 13:3441

42 Terumasa Ohkusa

truth of any statement does not in itself undermine his conviction thatpreviously built-up beliefs will again be accredited. Cross-examinationand dialogue may bring about agreement between opposite positions. Inprinciple, Socrates can commit himself to a certain content even whileit is temporarily suspended under elenctic examination.20

In this way, we can begin to appreciate that the attribution toSocrates of a total absence of belief is unsustainable in the light of hisphilosophical method. It would seem plausible thus far to equate anybelief that Socrates might hold in the course of an elenctic dialoguewith Vlastos’s weak knowledge, characteristic of humans, not the gods.But the fact that Socrates may profess a form of “knowledge” admittingof a weaker degree of certainty says nothing of the strength that mayattach to his convictions etc. of this kind; Socrates makes an evidentlymoral decision to lead his life in accordance with beliefs, judgments,interpretations, and not exclusively with knowledge. For example, hedeclares that he will never abandon positions he has taken up or hasfound assigned to him by his superior and which he believes (¹ghs£-menoj) best (28 d). Socrates expresses this commitment to the conse-quences of his beliefs even in the face of death or any major uncertainty.The contrast here is between the certainty attaching to Socrates’s inter-pretation (ò»qhn te ka� Øp�labon) of his philosophical task (28 e) andthe opacity of what follows death; the former certainty obliges Socratesto direct his life as he does.21 His life is based upon the principle that aphilosophical life in which one constantly examines oneself and othersis enjoined by divine command.22

Given Socrates’s commitment to the philosophical life, what he iscontrasting with divine knowledge in terms of the defining criterion ofcertainty must be the whole range of human beliefs, including but not

starting-point of any consensual discussion. This does not, however, deter So-crates from initiating a process of elenchus with his interlocutor.

20 This principle depends on a distinction between the examination of acertain proposition and its endorsement as a position. See e. g. Socrates in Plato’sGrg. 472 d – 475 e.

21 At the end of section 1 of this paper, we have seen the similar contrastbetween the certainty attaching to Socrates’s knowledge and the opacity of whatfollows death at 29 a–b. As far as the strength or certainty is concerned, I thinkthere is no clear distinction between knowledge and belief.

22 Note that these convictions of Socrates will hardly be verifiable. Neitherwill they be shared with his accusers. In my view, the absence of shareability inthis sense determines Socrates’s caution in avoiding the word ‘know’ here. ButSocrates’s cognitive terms – such as belief or conviction – may well admit of suchstrength as to be unwavering for Socrates. For Socrates’s positive convictions, seefurther 30 a 6 (o�omai), 35 d 7 (nom�zw), etc. See n. 23.

035_056_Ohkusa.pmd 11.05.09, 13:3442

43Socrates’s Avowal of Knowledge Revisited

limited to knowledge. Conversely, if we focus only on the conditions ofknowledge (like Vlastos’s two senses of knowledge), we miss the im-portance of other forms of human cognition that can make a Socraticlife steadfast.23

When Socrates makes definite assertions as to the value of livinglife in a particular way, he does not make any explicit distinction be-tween what he claims to know and what he strongly believes. The prob-lem of knowledge in these contexts necessarily concerns Socrates’s af-firmation to that effect – that is, that he knows. How do these assertionsor denials proceed? Socrates makes a firm denial of the knowledge at-tributed to a caricatured version of himself in Aristophanes’s play (19c–d). He also disowns the form of pedagogical knowledge peddled bySophists for cash (19 e – 20 e). Neither does Socrates profess the tech-nical knowledge of craftsmen he met (22 c–e).24 Whatever form ofknowledge Socrates lays claim to, it differs from the supposed knowl-edge of these practitioners in content, scope, practicality and level of

23 D. Wolfsdorf, “Socrates’ Avowals of Knowledge”, Phronesis 49 (2004) 89,139–140 (cf. Benson [n. 8] 84 n. 115, 227 n. 17), correctly points out that thepassage at 508 e 6 – 509 b 1 in Plato’s Gorgias – on the matters bound with chainsof iron and adamant, i. e. on the thesis that doing injustice is worse than sufferingit – only insists on Socrates’s belief that moral matters stand a certain way; that is,contra Vlastos, Socratic Studies (n.16) 59 n. 47, and “Socrates’ Disavowal ofKnowledge” (n. 16) 21 n. 48, he does not avow ethical knowledge here. Thus,however certain a proposition seems to Socrates, he is well aware (as in thispassage) of its disputable character when faced with those unsympathetic to him.

24 Socrates disqualifies notable politicians and poets from almost any knowledge(21 c – 22 c). On the other hand, Socrates finds some knowledge in craftsmen, asexpected (22 c–e); indeed, they are held to “know many fine things” (poll¦ ka�kal¦ �pistam�nouj) of which he himself is ignorant (ºp�stanto § �gë oÙkºpist£mhn). In this respect, they are wiser than Socrates (mou taÚtV sofèteroiÃsan), without ceasing to be vulnerable to the error of poets when they claimwisdom in “other, supremely important matters” (t«lla t¦ m�gista). Thesestatements offer another basis on which to take especial care in defining thecontent and context of Socrates’s knowledge-claims; even the “taÚtV” above maysuggest large divergences between forms of knowledge held by different parties.Woodruff (P. Woodruff, “Plato’s early theory of knowledge”, in: S. Everson [ed.],Companions to Ancient Thought 1. Epistemology [Cambridge 1990] 60–84)equates Socrates’s avowal and disavowal of knowledge with a distinction betweennon-technical and technical knowledge. This seems inadequate; even if restrictedto the field of non-technical knowledge or to ordinary language understandings,Socrates may avow some knowledge, as in the Apology (cf. also Euthd. 293 b–c,296 e – 297 a, Ion 532 d–e), but disavow other knowledge (unless he is claimingto be a polymath). This makes it difficult to uphold Woodruff’s distinction,especially when we come to consider Socrates’s avowal, rather than disavowal, ofknowledge. On a possible knowledge on Socrates’s part, see also my nn. 35, 36, 56.

035_056_Ohkusa.pmd 11.05.09, 13:3443

44 Terumasa Ohkusa

generality. Socrates’s disavowal of knowledge, then, may not amount toa sweeping renunciation of any knowledge involved, e. g. in pragmaticor technical activities. If Socrates hedges his disavowal of knowledge inthese terms, we may expect that he possesses some knowledge on cer-tain subjects or will claim some knowledge in a given respect.25

Socrates’s position may be compared with a sceptic’s more generaldisclaimer of knowledge. A sceptic like Arcesilaus would hesitate evento disavow knowledge for the same reason that he would be cautiousabout claiming it – he is reluctant to commit himself to any truthclaim.26 If so, a person’s definite disavowal of knowledge may well beregarded as a sign that he is ready, in principle, to avow some otherknowledge – in other words, that he credits the possibility of other de-terminate knowledges. As we saw previously, Socrates’s position is notthoroughly sceptical;27 even in the realm of morals, he is prepared defi-nitely to disclaim some ethical knowledge. Given that his disavowalsare not sweeping, it is reasonable to attribute to him a version of thequalified position suggested above: that he avows some knowledge inproportion to the strength of his other denials. Acceptance of this pic-ture of Socrates as differentiated from a certain kind of sceptic leads toa re-examination of the question of what exactly Socrates means by say-ing that human wisdom, including his, is worth little or nothing. Doesthis apparently categorical statement about human cognition not con-flict with the avowal of knowledge at 29 b describing his fortitude andway of life?

25 Wolfsdorf (n. 23) 75–142 attempts to look beyond Socrates’s evidentinconsistencies to get at the core of what Socrates – and behind him, Plato – reallybelieve about knowledge (interpretation [5]). But in Plato’s Gorgias Socratesexplicitly posits the consistency of an individual’s belief-set as pertinent to thetruthfulness or veridical nature of their assertions. Indeed, Socrates goes so far asto say that consistency is a distinctive feature of filosof�a (482 a–c). It thusbecomes problematic for Wolfsdorf to attribute Socrates’s deliberate shifts ofpositions to a “dramaturgical” strategy on Plato’s part, since by Socrates’s ownlights they would tend to make him unreliable, or a liar. For Wolfsdorf, Socrateschanges his positions so frequently because he wants to stimulate his interlocutorsto independent thought. The pedagogue Socrates, on this view, conceals his viewsfrom his interlocutors in order to educate them. But Socrates rather representshimself in his dialogues as a participant with his interlocutors in a search for truth,someone not already apprised of conclusions. If Socrates is genuinely usingdialogues to get at the truth, it would seem reasonable for him to stick with roughlyconsistent positions approximating his current grasp on the matter while makingadjustments e. g. for his interlocutor.

26 Cic. Ac. I, 44–45, cf. Woodruff (n. 24) 62.27 Cf. nn. 7–9.

035_056_Ohkusa.pmd 11.05.09, 13:3444

45Socrates’s Avowal of Knowledge Revisited

Socrates disclaims knowledge of the fine and good (21 b 4–5), andis well aware that he amounts to little or nothing in terms of wisdom (cf.23 b 2).28 However, these statements in themselves do not necessarilydisbar specific or partial knowledge. One plausible reading would be asfollows: because Socrates occupies a standpoint from which he cancompare human and divine knowledge, he might well see that the dif-ference is so great that he can declare the former practically nugatory.The wide gap between divine knowledge and characteristic humanmodes of cognition will prompt him to confess his ignorance withouthesitation.29 The possibility is, therefore, that his confession of igno-rance is made with reference to some degree of partial knowledgewhich Socrates can claim, which is however patently incomplete orungodlike. The knowledge attested at 29 b does not represent a goal inthe quest for moral judgment, but rather the starting-point of a processof serious enquiry, for example, concerning the nature of justice. If thisline of thought is possible, we may take Socrates’s avowal and dis-avowal of knowledge consistently: that is to say, as content-dependent(and of course, context-dependent, cf. pp. 41–42) rather than general.30

While Socrates cannot assert a synoptic and complete knowledge thatwill definitively answer the question what justice is (cf. R. I, 354 b–c),at 29 b he avows knowledge relative to the question of justice or injus-tice. When Socrates thinks of a wise soul in relation to some important

28 These passages (21 b 4–5, 23 b 2–4, 22 c 9 – d 1) might be thoughtproblematic, in that if we provisionally take Socrates to be asserting “I know thatI know nothing”, that would lead to the appearance of a logical paradox. On thispoint, some take the first “know” to refer to the comparatively weaker form ofcognition than knowledge (Stokes [n. 1] 19, 53; H. Tredennick, H. Tarrant, TheLast Days of Socrates [Harmondsworth 1993] 42, 44), while others suggest astrong epistemic claim as knowledge must be understood (Brickhouse & Smith [n.16] 33 n. 11, Woodruff [n. 24] 62 n. 3). In my view, choosing the weaker in-terpretation of cognition in the principal clause would implausibly undermineSocrates’s unwavering awareness of his own ignorance. As far as the strength ofthe first “know”, I agree with Brickhouse & Smith and Woodruff. As far as thepossible logical paradox is concerned, on the other hand, I agree with M. F. Burnyeat,“Antipater and Self-Refutation: Elusive Arguments in Cicero’s Academica”, in:B. Inwood, J. Mansfeld (eds.), Assent and Argument: Studies in Cicero’s AcademicBooks (Leiden 1997) 291, that in the Apology what Socrates declares is morenuanced than “I know that I know nothing”. Cf. pp. 43–44.

29 Cf. Vlastos, Socratic Studies (n. 16) 62; idem, “Socrates’ Disavowal ofKnowledge” (n. 16) 28; idem, Socrates: Ironist and Moral Philosopher (Cam-bridge 1991) 221 n. 76; idem, “Happiness and Virtue in Socrates’ Moral Theory”,Topoi 4 (1985) 19 n. 69; I. Park, “Inner Action”, Bulletin of Kansai PhilosophicalAssociation 24 (1989) 81, Kraut (n. 14) 272–274.

30 Cf. Park (n. 29) 81; Nozick (n. 46) 143–145. Cf. also Guthrie (n. 1) 88.

035_056_Ohkusa.pmd 11.05.09, 13:3445

46 Terumasa Ohkusa

matter, he accesses some concept of the former knowledge, and so iscompelled to profess his ignorance. It could be tentatively concludedthat for Socrates to avow partial knowledge is one thing and to claimcomplete or synoptic knowledge quite another. In other words, the as-sertion of knowledge at 29 b is far from answering the question whatjustice is.31

In conclusion, I will briefly summarise the chief points made in thissection. Vlastos’s weak knowledge characterised as fallible and uncer-tain does not explain 29 b well. This is because the knowledge in ques-tion rather seems strong and secure especially because of the strikingcontrast between Socrates’s avowal of knowledge at 29 b and people’sdelusion, ¢maq�a. Further, Vlastos’s two senses of knowledge say noth-ing of Socrates’s strong conviction. What should be compared with di-vine knowledge in the light of certainty in Vlastos’s sense is probablythe whole range of human cognition, not restrictively human knowledge(pace Vlastos). This is because Socrates appeals to both knowledge andbeliefs to lead his unwavering philosophical life. (We should take bothknowledge and beliefs as they are in the texts: it is unnecessary to as-similate some beliefs with knowledge, or some knowledge with be-liefs.) My alternative interpretation is that Socrates can avow or dis-

31 It would be natural at this point to open out the argument into a considerationof the so-called Socratic fallacy, referring to Socrates’s supposed insistence on thepriority of the definition of objects to any attestation of knowledge concerningthem. Let me just say for now that I do not judge Socrates’s reasoning to befallacious. As I understand it, Socrates does not demand for objects of knowledgeto be defined in anything other than a context-dependent way (cf. A. Nehamas,“Socratic Intellectualism”, Proceedings of the Boston Area Colloquium in AncientPhilosophy 2 [1986] 285–293 and idem, The Art of Living [Berkeley etc. 1998]217–218 n. 61); indeed, he tends to ask his interlocutors to define terms aboutwhich they have previously professed some especial competence e. g. as a teacher.Moreover, it is often indicated that the priority of definition at 6 d–e in theEuthyphro may represent for Socrates a sufficient, or completely satisfactory,rather than a necessary condition for the judging of examples (pace P. T. Geach,“Plato’s Euthyphro: Analysis and Commentary”, Monist 50 [1966] 370–372. Onhow to read this passage of Euthyphro, I agree with Nehamas, “Socratic Intel-lectualism”, 275–293, and Brickhouse & Smith [n. 16] esp. p. 49. IncidentallyBenson [n. 8] 120 n. 33 puts aside Nehamas’s interpretation): cf. G. X. Santas,“The Socratic Fallacy”, Journal of History of Philosophy 10 (1972) 136; idem,Socrates (London 1979) 116; Kraut (n. 14) 209 n. 38; Vlastos, “Socrates’ Disavowalof Knowledge” (n. 16) 23 n. 54; idem, “Is the ‘Socratic Fallacy’ Socratic?”,Ancient Philosophy 10 (1990) 7; J. Beversluis, “Does Socrates Commit theSocratic Fallacy?”, American Philosophical Quarterly 24 (1987) 211–223; M. McPher-ran, The Religion of Socrates (Pennsylvania 1996) 180 n. 11; Ch. Kahn, Plato andthe Socratic Dialogues (Cambridge 1996) 157, 181–182. Cf. Benson (n. 8) 144–147.

035_056_Ohkusa.pmd 11.05.09, 13:3446

47Socrates’s Avowal of Knowledge Revisited

avow knowledge in a content- and context dependent manner. For ex-ample, Socrates avows knowledge at 29 b, but he might well refrainfrom claiming it at certain junctures, e. g. in front of Thrasymachus,who would contest 29 b. Finally we saw that Socrates’s disavowal ofknowledge is not sweeping. Thus it can reasonably be expected that heclaims some qualified knowledge, e. g. on a certain subject or in a givenrespect. Then the problem would be how we should interpret the seem-ingly categorical disavowals of knowledge that Socrates sometimesmade. Actually this is where we should appeal to the distinction be-tween human and divine knowledge, not regarding certainty/uncertaintyas Vlastos did. Since Socrates occupies a standpoint from which he cancompare human and divine knowledge, he can declare his partialknowledge practically nugatory. However, this is compatible withSocrates having some knowledge.

3. The scope of Socratic knowledge

Socrates sets great store by the elenchus as a method of arriving attruths upon which he and his interlocutors can agree. The elenchus initself, as he understands it, partakes of the character of a search fortruth. Socrates expresses delight with his encounter with Callicles,comparing him to an excellent touchstone (Grg. 486 e 5–6). Why isSocrates so sure about the truth-value of an agreement? Generally,agreements seem subject to instability: they depend on whether the in-terlocutors are skilled, whether the procedure of a dialogue is suitablycarried out, etc. Socrates pays close attention to these issues in attribut-ing truth to agreement.32 Callicles is praised for those qualities that areconducive to his reliability as a discussant: knowledge, goodwill andfrankness (487 a 2–3 �pist»mhn te ka� eÜnoian ka� parrhs�an).These three credentials are taken to be necessary for the elenchus toorient itself effectively towards the truth. Socrates thus addressesCallicles (487 e 6–7): “In reality, then, agreement between you and me

32 Socrates sometimes doubts his own qualifications as a searcher for truth(Chrm. 175 e 5 – 176 a 1, R. I, 336 e 10) and frequently refers to the tentativenessof his conclusions (Grg. 480 a 1, 480 b 3, 480 e 3–4 etc.). The elenchus affords itsparticipants freedom to take back their assent to any previous proposition (Chrm.164 d, 165 a–b, Grg. 462 c 4, 464 a 1, 506 a 4–5, cf. ¢naq�sqai at 461 d 3 and 462a 3. Cf. I. Park, “Rhetoric, Persuasion, and Dialectic”, Journal of Classical Studies47 [1999] 105–106; Vlastos, Socratic Studies [n. 16] 17 n. 51; idem, “The SocraticElenchus”, Oxford Studies in Ancient Philosophy 1 [1983] 44–45 n. 47 contraT. Irwin, Plato’s Moral Theory [n. 11] 39).

035_056_Ohkusa.pmd 11.05.09, 13:3447

48 Terumasa Ohkusa

(¹ �m¾ ka� ¹ s¾ Ðmolog�a) will then obtain the goal of the truth (t�loj... tÁj ¢lhqe�aj)”.

On the other hand, Socrates converses with anyone, young or old(Ap. 30 a 3, 9) who is willing (Ap. 33 a 7, b 3). He even indicates that thesolution of the matter under examination will be a common good for all(cf. Chrm. 166 d 5, Grg. 505 e 6). Socrates would seem to postulate thatno controversy over truth obtains among people at the deepest level.33

When Polus, in the Gorgias, laughs in Socrates’s face, Socrates returnsto first principles in the belief that his views and those of his interlocu-tor will ultimately be found to coincide:

For I think that you and I – and also the other people – believe thatdoing injustice is worse than suffering it, and that not paying justice isworse than paying it (Grg. 474 b 2–5).

In highlighting such passages, Vlastos suggests that some version ofthe theory of recollection might be supposed in the Socratic elenchus,though in the end these represent distinct methods for Vlastos.34 In thepresent case, Socrates attributes to Polus beliefs that are the opposite ofPolus’s current assertions on the basis that these beliefs are implicitlyentailed by other of Polus’s beliefs. In my view, a similarly structuredassertion is made by Socrates in the Apology, a work generally held tobe no chronologically closer to the Meno than the Gorgias.35 When

33 Cf. Woodruff (n. 24) 79–80.34 Throughout his Socratic studies, Vlastos understands the elenchus, in its

proper usage, as a procedure used by Socrates to refute his interlocutors’ statementsor to correct such in the light of his interlocutors’ other beliefs. It is not until theintroduction of a positive theory of recollection that Socrates finds a methodcapable of adducing proofs. Socrates’s attestation of having proved the truth in theGorgias is thus extraordinary for Vlastos, meriting close examination. Vlastosprovocatively refuses to accept the famous passages of geometrical demonstrationin the Meno as examples of the elenchus: Socratic Studies (n. 16) 5 and “Socrates’Disavowal of Knowledge” (n. 16) 32; for Vlastos, Plato retrenches from theelenchus when the theory of recollection is introduced. This developmental line ofreading has attracted the support of a number of scholars, from whom I dissent. Cf.also Vlastos (n. 29) 118–119; idem, “Elenchus and Mathematics: A Turning-Pointin Plato’s Philosophical Development”, AJPh 109 (1988) 373.

35 I elsewhere suggest my grounds for reservation from scholars who distinguishPlato’s earlier and middle dialogues according to the strictness with which theyapply the priority of knowledge-what, representatively found clearly in Kraut (n.14) 274–277 (interpretation [4]), and also in Vlastos “Socrates’ Disavowal ofKnowledge” (n. 16) 26 n. 65; idem, Socratic Studies (n. 16) 71 n. 14; Beversluis(n. 31) 218, 221 n. 4, etc. This interpretation both over-schematizes Plato and findssome difficulties in the texts: in the Meno, Socrates may possess some knowledge

035_056_Ohkusa.pmd 11.05.09, 13:3448

49Socrates’s Avowal of Knowledge Revisited

Meletus’s criticisms of Socrates converge in an accusation that Socratesdoes not admit the existence of god at all, Socrates says (Ap. 26 e 6–7):

You are not credible, Meletus, and that, it seems to me, not even toyourself.

Again Socrates denies that Meletus can plausibly hold views con-trary to his own.

In either case, Socrates’s comment seems to suggest that people canhold some ideas latently which contradict their own statements (cf. Grg.495 e 1–2).36 On this point, Vlastos attributes the following assumptionto Socrates: anyone who ever has a false belief will always have at thesame time true beliefs entailing its negation37 (I call this Vlastos’s as-sumption A). Thus anyone who bears false ideas will tacitly accept thenegation of their original opinion.38 I agree with Vlastos that Socrates’selenctic procedures are predicated on Socrates subscribing to his as-sumption A. Even when Socrates initially seems to dismiss an inter-

as implied at 75 d 5–7 (proomologÍ, �rwtîn in my reading, following R. S. Bluckand E. S. Thompson respectively), 85 d 9–10, 98 b 1–5 etc.

36 Another early passage with possible affinities with the so-called theory ofrecollection is the miscommunication between Socrates and Gorgias at 459 c – 460a in Plato’s Gorgias – miscommunication because Socrates’s formulation here, ina wording possibly reminiscent of recollection, i. e. proepist£menon (459 e 1–2),proeidÍ (459 e 7), does not for a moment consider that Gorgias could teach hispupils the nature of e. g. the good, bad, beautiful, ugly, just, unjust etc.; Gorgiashimself maintains that his pupils will pick these up as readily as they absorb hislessons in rhetoric (460 a 3–4). In asking how Gorgias’s pupils could come toknowledge of these subjects, Socrates appears to rule out a normal learningprocess of knowledge passing from a teacher to a pupil, instead apparently restingon an idea of introspection closer to recollection. If we limit ourselves, however,to observing the content- and context-dependent nature of Socrates’s assertions ofknowledge, we will not have to reckon with the same difficulties as developmentalistaccounts.

37 Perhaps worrying over whether the beliefs in question can be accounted‘true’, Vlastos drops the predicative adjective ‘true’ (Socratic Stidies [n. 16] 56and “Socrates’ Disavowal of Knowledge” [n. 16] 18), adopting elsewhere theexpression ‘true beliefs’ (Vlastos, Socratic Stidies [n. 16] 25; [n. 32] 52; Socrates[n. 29] 114; [n. 34] 369).

38 Cf. Vlastos, Socratic Studies (n. 16) 23–24; idem (n. 32) 51. On tacit (covert)beliefs, Vlastos has recourse to this example: one can tacitly believe that triangles’interior angles sum to two right angles without necessarily having formed (inchildhood, for instance) any explicit (overt) idea on the matter. Vlastos describesSocrates’s belief concerning the implicit relationship of negation between people’sfalse and true opinions as constituting a “tremendous assumption” on Socrates’spart (Vlastos, Socratic Studies [n. 16] 25; [n. 32] 52).

035_056_Ohkusa.pmd 11.05.09, 13:3449

50 Terumasa Ohkusa

locutor’s case out of hand like those passages above, which are un-usual,39 he remains concerned to evaluate which side of the argument –his or his opponent’s – appears more persuasive to both parties: “Thenyou’ll answer?” (Grg. 474 c 1) “Enquire, then, along with me, gentle-men …” (Ap. 27 a 9). In these cases, his expectation is that his inter-locutor’s stated view will fail to be reasonable in that it will prove in-consistent with his other avowed beliefs. In teasing these out, Socratescan thus obtain a series of agreements on a number of successive mat-ters (Grg. 474 c – 475 e, Ap. 27 a – 28 a). Socrates needs to turn hisinterlocutors into witnesses of the successive phases of his arguments’proof so long as their assent remains essential to him in the establishmentof the truth.40 Socrates’s deployment of the elenchus as an effectivemethod of ascertaining truth would seem to depend on his endowing hisinterlocutors with some knowledge or understanding, in the sense thatthey are discerning enough to judge what is true and what is false.

Nevertheless, for Vlastos, there remain philosophical problems inSocrates’s positing of elenctic dialogues as a mechanism for determin-ing truths. This is because dialogues cannot by themselves supply anycriterion for determining whether a belief is true; rather, they merelydemonstrate the incoherence of interlocutor’s belief-sets. But if Vlastosis correct in attributing his assumption A to Socrates, there is a basis onwhich Socrates becomes entitled to treat the agreed propositions emerg-ing from dialogues as having a certain truth-value. This is becauseSocrates is seen as proposing true statements and his interlocutors,whatever their earlier assertions or other beliefs, as being competent toendorse them. In other words, in conversation, Socrates is likely to steerhis discussions in the direction of shared truth, through examining theentirety of the beliefs of his interlocutors. According to Vlastos, Socratesplaces a greater degree of faith in the elenchus than is warranted by its

39 These passages seem exceptional against the backdrop of Socrates’s character(as indicated in p. 41 and n. 19).

40 On this point, see Vlastos, Socratic Studies (n. 16) 20–21; (n. 32) 48, 53.Vlastos states that Socrates typically makes his partners witnesses of his ownviews; the textual evidence he cites, though, shows something else: the fact thatthe elenchus typically proceeds through Socrates’s gaining the assent of oneinterlocutor alone (Grg. 474 a 5–6). Moreover, it is clear that Socrates judges theelenchus worthwhile even should he fail to carry his interlocutors – that is, bothhis refutations of others and crucially their refutation of him yield a cognitivebenefit (Ap. 22 a 7–8, Grg. 458 a 1–b 1, 470 c 6–7, 506 c 1–3). Vlastos’squotations (Vlastos, Socratic Studies [n. 16] 21–27, T 21–24; [n. 32] 48–54, T 22–25) tend to illustrate, pace Vlastos, Socrates’s confidence in the elenctic method,rather than showing that dialectical situations as such are advantageous to Socrates.

035_056_Ohkusa.pmd 11.05.09, 13:3450

51Socrates’s Avowal of Knowledge Revisited

merely serving as the formal guarantor of its conclusions. The criterionof truthfulness of a conclusion, though, is not that it is finally subject toagreement, but that it is Socrates who proposed it. Vlastos argues thatSocrates’s experience of debating, in other words, inductive evidence,leads him to the position that his own belief-set consists exclusively oftrue beliefs.41 Socrates therefore becomes confident in his own abilitiesas an arguer to determine the truth. It follows for Vlastos’s line ofthought that such a formidable debater might be in no need of furtherenquiry, obviating the need for any interlocutor or even dialogue.42

So, for Vlastos, Socrates must be apprised of the truth of those as-sertions he submits to elenctic conversation. For Irwin, however,Socrates’s title to this kind of knowledge comes into doubt, as Socrates,like anyone else, lacks a criterion by which he might account his truebeliefs concerning any object to be “knowledge”. Now, as Irwin sug-gests, Socrates may remain convinced of the truth of his beliefs on anysubject of discussion, without attaching to them any explicit claim toknowledge.43 Yet Irwin may be hasty in having Socrates separate truthand knowledge so readily; since when Socrates says that he cannot at-tend a claim with knowledge, he surely means that he is unable to an-swer disputes as to its truth. Whether Socrates has true beliefs or not,and which these are, must remain a matter for independent examination.

The prevailing literature on this topic would accept as uncontentious theidea that the Platonic Socrates holds some true beliefs. But which? ForNozick, Socrates, in disavowing knowledge of F, necessarily holds himself tolack true belief in the matter of F. Yet without claiming to define F, it may stillbe possible, via the process of an elenchus, to begin considering, and to seekto know, any topic insofar as it can be identified as pertaining to the nature ofF.44 For both Vlastos and Irwin, the elenchus is a negative procedure restingon Socrates’s convictions and concerned to expose falsehoods.45 But rather

41 Vlastos, Socratic Studies (n. 16) 28; (n. 32) 55.42 In my view, Vlastos here goes too far in failing to capture the extent to

which Socrates’s discussions represent to him genuine opportunities to thrash outunsuspected and unformulated truths.

43 Irwin (n. 11) 40–41.44 This would understand the so-called priority of definition as applying less

restrictively than it is sometimes imagined. On the context-dependency of thepriority of definition, see n. 31. On this point, M. F. Burnyeat, “Examples inepistemology: Socrates, Theaetetus and G. E. Moore”, Philosophy 52 (1976) 389;Nehamas, “Socratic Intellectualism” (n. 31) esp. p. 290, would seem to hold acontext-dependent interpretation. Cf. n. 31.

45 For those writers, Plato’s positive philosophical contributions may be foundin or extrapolated from statements he arrived at independently of dialogue.

035_056_Ohkusa.pmd 11.05.09, 13:3451

52 Terumasa Ohkusa

than understanding Socrates as directing discussions on the basis of prin-ciples not subject to examination (like Irwin), we can see Socrates as movingfrom point to point guaranteed by agreement between himself and some ex-pression of his interlocutors’ minds. In putting this point, Nozick modifiesVlastos’s assumption A as to the latency of true beliefs for Socrates in humancognition.46 Ceasing to view the elenchus as a demonstration, Nozick’s newassumption is that Socrates47 and his interlocutor are on a par, in the sensethat some knowledge is attributed to both parties.48 In order to endorse theSocratic emphasis on the value of mutual agreement, so that the elenchusmay be regarded as a search for truth, we will then admit the substantialimportance, not of those argumentative phases in which Socrates professeshis true beliefs, but of those in which both parties come to believe somethingtrue or know something to be true.49

While Socrates strictly denies knowledge to himself and others, henonetheless engages in dialogues which he is willing to construe as thepursuit of truth; moreover, he tries his utmost to offer himself and oth-ers the possibility of knowledge. Callicles’s personal qualities were ad-mitted as factors potentially orienting the elenchus towards truth.50

46 R. Nozick, “Socratic Puzzles”, Phronesis 40 (1995) 149–150.47 According to Nozick, the excellence of Socratic knowledge is that he knows

that he does not know (what F is). Nozick describes what are elsewhere calledSocratic doctrines (fragmentary statements of Socrates’s seeming to attest toconsistent positions e. g. it is better to suffer injustice than to commit it) asSocrates’s knowledge. At this point Nozick also takes care to limit the scope ofSocratic ignorance. Not knowing what F is is for Nozick compatible with knowingother things about F. Cf. n. 30.

48 Nozick’s modification concerns only this point; what Vlastos calls people’s‘true beliefs’ i. e. the beliefs that they will hold onto when confronted with achoice in the course of the elenchus, Nozick calls ‘knowledge’. This modificationsuggests that both Socrates and his interlocutors have some tacit knowledge (cf. n.38). Indeed, for Nozick, many of people’s beliefs are likely to be true, because‘knowledge’ has more stability (tenacity, stickiness) than false beliefs, which tendto be rejected after reflection or experience. Nozick’s argument also admits thepossibility of people’s arriving at knowledge other than through the elenchus (cf.Nozick [n. 46] 151).

49 The implication here would be that, in principle, the truth of the conclusionof an elenctic discussion does not entail the truth or reliability of elenctic premises.In other words, the elenchus does not comprise a method of logical induction. It ispossible on a case-by-case basis for Socrates to rid himself of a commitment to hispremises – this is a point on which I agree with Benson (n. 8) 47–52 and M. Nakahata,“Truth in Dialogues – A Note on the Socratic Elenchus (II)”, Methodos 29 (1997)11–20.

50 Socrates later suggests that Callicles’s goodwill was lacking over the courseof their discussion (Grg. 499 c).

035_056_Ohkusa.pmd 11.05.09, 13:3452

53Socrates’s Avowal of Knowledge Revisited

Socrates is presumably willing to posit those three qualities in every-one, at least at the latent level. If the elenchus is to work well, it rests onthe assumption that interlocutors have some prior beliefs or knowledgein an implicit or explicit form, which will be adopted as a commonground of truth and tested over the course of time.51 Indeed, this mustbe a necessary, considering, with Vlastos, that the elenchus, in point oflogic, can be said at most to indicate inconsistencies of beliefs.52 Sosome postulate of people’s capacity for forming true concepts wouldappear a prerequisite of constructing the elenchus as a search fortruth.53 As long as what Socrates is searching for is not true belief butknowledge,54 his dialogical procedure must be premised on the assump-tion that speakers (including himself) possess some ethical knowledge.

In the conversation with Callicles, Socrates allegorically representsa person’s awareness of his own ignorance through the figure of a sea-pilot. The sea-pilot is a man of modesty and humility despite saving

51 In recent Stoic studies, Anthony Long upholds a version of Vlastos’sassumption A as underlying not only Socratic but also Epictetan dialogical strategy(A. A. Long, Epictetus: A Stoic and Socratic Guide to Life [Oxford 2002] 82–83).

52 In my view, it is only when some truth is attributed to us in advance –whether in tacit or explicit form – that we can justifiably infer from the ‘consistency’of our beliefs to their “truth”. This postulation of latent truth goes some way toexplaining why Socrates associates a person’s consistency with the truth of theirprofessions. He responds to Callicles’s changes of position in argument bypredicting that Callicles will be at discord with himself throughout his whole life(Grg. 482 b–c) – implicitly using consistency as a measure of truth and falsehood. Itis important to note that Socrates characteristically examines the truth of propositionsnot in isolation or according to their purely formal aspect (as p, q, r, etc.), butinsofar as they fit or not with the rest of a person’s belief-set, as evidenced by theentirety of his person. Indeed, for Socrates, it would be hopelessly arbitrary toselect one of his interlocutors’ propositions as preferable to any other withoutpursuing further substantial indications as to propositions’ truth-value. Socrates’sbias in favour of personal and argumentative consistency partly solves the structural“problem of the elenchus” as set out by Vlastos: that is, why Socrates seems tobelieve that he has refuted a proposition (p) as false when logically he has done nomore than demonstrate its incompatibility with other propositions (q, r, s, etc.).

53 In this way, the refusal of Socrates’s adversaries Meletus and Callicles toanswer his questions suggests that to some degree they are persuaded of the truthof Socrates’s arguments (Grg. 513 c 4–6, 517 c 7 – d 5, 518 a 5–7, Ap. 24 d 7, 27 c10). This raises the question of why Socrates does not always carry his interlocutors(e. g. in the Apology acquit himself). Socrates himself explains this, especiallypointing to their love of demos (Grg. 481 d 5, 513 c 7), their slander, grudge,anger, stubbornness (Ap. 31 a 3–5, 34 c 7 – d 1), the time restrictions imposed onhim (Grg. 455 a 2–7, 513 c 8 – d 1, Ap. 19 a 2, 36 d 5–6, 37 a 8 – b 2, 38 c 1), etc.

54 Cf. Grg. 472 c 6 – d 1, 505 e 4–5.

035_056_Ohkusa.pmd 11.05.09, 13:3453

54 Terumasa Ohkusa

people’s “bodies and property” as well as their “lives” from perils. Henever assumes an air of importance by conceiving his work to be spe-cial. Socrates says of him (Grg. 511 e 6 – 512 a 2):

For I suppose he knows (�p�statai) how to reason (log�zesqai) thatit’s unclear (¥dhlon) which passengers he has benefited by not lettingthem drown, and which ones he has harmed;55 he knows (e�dèj) hehas put them ashore no better than they were when they boarded,either in body or in soul.

It is possible that such a fine description and appreciation of the sea-pilot is available only to one who grasps the nature and scope of knowl-edge as subtly as Socrates himself.56 Socrates posits a limit to what thesea-pilot knows – and to that extent, his emphasis falls on the negativeside of the sea-pilot’s cognition.57 At the same time he suggests, how-ever, that the sea-pilot is both skilful and acute in grasping the moralconsequences of his actions. Socrates depicts the sea-pilot as knowing“how to reason” and as having formed ideas of various subjects. Thiscognitive attainment represents a form of knowledge. By being awareof one’s own ignorance, like the sea-pilot, a person positions himself asoriented towards truth – for example, towards the horizon of the signifi-

55 Similarly, in the Apology, Socrates claims that he cannot know which of hisdiscussants he has helped or strengthened, and which he has harmed, throughengaging them in elenctic dialogue (cf. Ap. 30 b 4–6, 33 b 3–6).

56 Vlastos, Socratic Studies (n. 16) 47, 59 n. 47, and “Socrates’ Disavowal ofKnowledge” (n. 16) 10, 21 n. 48, also adduces a different passage about the sea-pilot (Grg. 512 b 1–2) to a consideration of the basis for Socrates’s avowal ofknowledge. The analogy of the sea-pilot itself does not support Socrates’s knowledge-claim, but rather suggests a possible prototype for a form of knowledge. SeeR. Bambrough, “Plato’s Political Analogies” (1971), reprint. in: G. Vlastos (ed.),Plato: A Collection of Critical Essays II (Notre Dame 1978) 187–205, for theimportance of the sea-pilot analogy to the Republic. It is also worth rememberingthat Socrates prides himself on his expertise on politics in its proper sense at 521 din the Gorgias.

57 Wolfsdorf and I differ considerably in our readings of Socrates’s analogybetween the philosopher and a sea-pilot. For Wolfsdorf, the sea-pilot stands as oneof a series of figures, including Socrates’s interlocutors Callicles, Euthyphro, Hippias,and so on, whose views are ultimately refuted and whose claims to knowledgeshown to be illegitimate. In my reading, the sea-pilot is a figure of Socraticknowledge (especially in the sense that his knowledge is related to a consciousnessof his own ignorance). In Plato’s dialogues, philosophers are compared to sea-pilots in a positive sense: while both may be caricatured as star-gazing babblers,both train their faculties on abstract objects of contemplation (compare R. VI 488d – 489 a with Phdr. 269 e – 270 a, especially with regard to ¢dolesc�a ka�metewrolog�a).

035_056_Ohkusa.pmd 11.05.09, 13:3454

55Socrates’s Avowal of Knowledge Revisited

cance of the pilot’s rescues as such (“which passengers he has bene-fited, and which he has harmed”). Socrates’s, and the sea-pilot’s, ques-tions paradoxically arise out of their scrupulously envisaged, or delim-ited, understanding of a world of which they disclaim knowledge.

Socrates always envisions the fine and good in every dialogue, andit is this viewpoint that stimulates him to ask the question what x is.Socrates then searches for a complete and synoptic knowledge capableof answering that enquiry. The question is predicated on his having al-ready obtained some degree of conviction, or possibly partial knowl-edge, such that he may require a complete and synoptic account of theconditions by which his cognition may be accounted knowledge in afull sense.58 Socrates’s whole manner of life is already supported by hisbroad vision before he initiates an elenchus with any interlocutor. Inthis sense, his life appears consistent with itself, and his words andbehaviour stable. Socrates’s careful delimitation of his knowledge inthe midst of his wider ignorance is comparable to the modest claims ofthe good sea-pilot. It would be natural and reasonable to assume some-thing parallel here: Socrates’s broadly-conceived cognition is likewiseso rich that he can avow knowledge clearly, at 29 b in the Apology,while at the same time remaining deeply conscious of his own igno-rance in the moral sphere.*

Terumasa Ohkusa

Kyoto University

 òå÷åíèå íåñêîëüêèõ ïîñëåäíèõ äåñÿòèëåòèé èññëåäîâàòåëè ïëàòîíîâ-

ñêîé ýïèñòåìîëîãèè âåäóò äèñêóññèè î òîì, â êàêîì ñìûñëå Ñîêðàò

ïðèçíàåò è â êàêîì íå ïðèçíàåò çà ñîáîé îáëàäàíèå çíàíèåì. Â ñòàòüå,

58 On the relation between “what is x?” question and its synoptic character,see esp. Euthphr. 6 d 9 – e 7, R. I, 354 b 9 – c 1, Men. 71 b 3–4. Socrates considersthat the answer to “what is x?” should cover all its instances without exception,coextensively.

* This paper is a revised version of “Socrates’ Avowal of Knowledge”,published in Methodos 32 (2000) 21–32 and written in Japanese. I wish tomake grateful acknowledgement of a generous grant from the Japan Society forthe Promotion of Science which enabled a prolonged stay in Cambridge duringwhich I prepared the paper. I am very much indebted to David Sedley, GeoffreyLloyd and Robert Wardy for their constructive comments and suggestions on anearlier draft. I also owe special thanks to the thorough review by AlexanderVerlinsky, which prompted revision of this paper.

035_056_Ohkusa.pmd 11.05.09, 13:3455

56 Terumasa Ohkusa

íàðÿäó ñ êðèòèêîé ïðåæíèõ èíòåðïðåòàöèé (Ã. Âëàñòîñà, Ò. Èðâèíà, Ã. Áåí-

ñîíà è äð.), ïðåäëàãàåòñÿ ðåøåíèå ýòîé ïðîáëåìû, îñíîâûâàþùååñÿ â

ïåðâóþ î÷åðåäü íà èíòåðïðåòàöèè ïàññàæà èç ïëàòîíîâñêîé Àïîëîãèè

Ñîêðàòà (29 b), à òàêæå íà àíàëèçå ñîêðàòîâñêîãî ìåòîäà èññëåäîâàíèÿ

(�legcoj). Àâòîð äîêàçûâàåò, ÷òî ñëåäóåò ïðèíèìàòü âñåðüåç êàê óòâåðæ-

äåíèÿ Ñîêðàòà â “Àïîëîãèè”, ÷òî îí îáëàäàåò çíàíèåì, òàê è îòðèöàíèÿ

ýòîãî, êàæóùååñÿ æå ïðîòèâîðå÷èå ìåæäó íèìè ðåøàåòñÿ â òîì ñìûñëå,

÷òî ïîäîáíûå óòâåðæäåíèÿ çàâèñÿò âñÿêèé ðàç îò êîíòåêñòà è îò ñîäåð-

æàíèÿ çíàíèÿ, êîòîðîå Ñîêðàò èìååò â âèäó. Óòî÷íåíèÿ, îòíîñÿùèåñÿ ê

ýëåíêòè÷åñêîìó ìåòîäó, ïîçâîëÿþò äàëåå îïðåäåëèòü, â êàêîì ñìûñëå

Ñîêðàò ïðèçíàåò, à â êàêîì íå ïðèçíàåò çà ñîáîé îáëàäàíèå çíàíèåì.

 çàêëþ÷åíèå âûäâèãàåòñÿ ïðåäïîëîæåíèå, ÷òî ôèãóðà êîðì÷åãî â “Ãîð-

ãèè” (511 e � 512 a), îáëàäàþùåãî çíàíèåì, êîòîðîå ñòðîãî îãðàíè÷åíî

îïðåäåëåííîé ñôåðîé, íî â íåé áåçóñëîâíî íàäåæíî, ñëóæèò àíàëîãèåé,

ïðîÿñíÿþùåé òîíêî äèôôåðåíöèðîâàííóþ ïîçèöèþ ïëàòîíîâñêîãî Ñî-

êðàòà â ýïèñòåìîëîãè÷åñêèõ âîïðîñàõ.

035_056_Ohkusa.pmd 11.05.09, 13:3456